The effects of elevated pain inhibition on endurance exercise performance

Participants

To control for potential gender (Popescu et al., 2010) and age (Grashorn et al., 2013) related variations in pain inhibitory capacity, 12 healthy, recreationally active males between 18 and 40 years of age were recruited for participation. All participants were recruited from the local University student population. Participants were not engaged in regular strength training programs or competitive sports predominantly involving knee extensor muscles (e.g., cycling, weightlifting). They were also free from symptoms of chronic pain and did not present with contraindications to tDCS (see Villamar et al., 2013).

Participants were instructed to refrain from consuming caffeine or analgesic medications or engaging in vigorous physical activity for at least 24 h prior to the experimental sessions. Adherence to these instructions was confirmed on arrival.

In Cogiamanian et al. (2007), changes in endurance performance following tDCS were assessed across active anodal (n = 9; M =  − 21.11 ± 5.52), active cathodal (n = 9; M =  − 35.77 ± 3.39) and control (n = 15; M =  − 39.33 ± 3.32) conditions. A-priori power analysis based on data presented in Cogiamanian et al. (2007) indicated a total sample of eight participants was needed to uncover a statistically significant difference in endurance time between active and sham conditions with a statistical power of 0.80 at an alpha level of 0.05.

Pain inhibition assessment

In the current study, pain inhibitory capacity was assessed using a sequential CPM design, whereby a noxious test stimulus (TS) is presented before and immediately after a noxious conditioning stimulus (CS) is applied to a spatially remote bodily region. This method is in contrast to the simultaneous or parallel CPM design where the second TS and the CS are presented simultaneously. The sequential design was preferred as it has been suggested to provide a clearer indication of endogenous pain modulation (Yarnitsky et al., 2015).

Pressure pain threshold (PPT) assessment was used as the TS. Using a manual pressure algometer (Wagner Force Dial FDK 20, Wagner Instruments, Greenwich, CT), pressure was gradually applied (approximately 1 kg/s) to the dorsal surface of the index finger of participants’ non-dominant hand. Participants were instructed to verbally indicate when the pressure first caused pain, at which point the pressure was released and the peak force required to elicit pain was recorded. After 30 s, this assessment was repeated, allowing for an average of the two trials to be calculated. Participants wore a blindfold throughout PPT assessment so as to minimise potential distraction. The use of PPT assessment as the TS in the CPM protocol is well accepted (see Pud, Granovsky & Yarnitsky, 2009).

Cuff occlusion of the non-dominant arm was used as the CS. A sphygmomanometer cuff was placed approximately 2 cm proximal to the cubital fossa and gradually inflated (∼20 mmHg/s) until participants first reported pain. The cuff then remained at this level of inflation while participants verbally rated their perceived pain level at 10 s intervals using a 0–10 numerical rating scale (0–10_NRS). If reported pain remained stable for two consecutive intervals, or if ratings of perceived pain decreased, then the cuff was inflated by ∼20 mmHg. Once a perceived pain intensity of 5 on the 0–10_NRS was reported, the cuff was fully deflated. Cuff occlusion has previously been shown to be an effective CS for the assessment of pain inhibitory capacity (Cathcart et al., 2009).

Immediately following the removal of the CS, PPT was assessed again. As recommended by Yarnitsky et al. (2015), CPM was calculated by subtracting post-conditioning PPT from pre-conditioning PPT. Therefore, lower CPM values represent enhanced pain inhibitory capacity.

Force measurement for exercise performance trials

An isokinetic dynamometer (Humac NORM, Computer Sports Medicine, Stoughton, MA) was used to assess force production during maximal voluntary contraction (MVC) and endurance tasks. Participants sat in an upright position perpendicular to the dynamometer while adjustments were made to ensure the correct alignment of the dynamometer’s axis of rotation with the lateral femoral epicondyle of the non-dominant leg. To limit movement during force production, participants were fastened to the chair at the shoulders and hip as well as the thigh of the involved leg. Throughout both performance tasks, force was applied by the non-dominant leg, fixed at 90° of flexion, through the padded arm of the dynamometer fastened slightly above the medial malleolus.

Maximal voluntary force production

Three submaximal familiarisation trials were performed 60 s prior to the MVC trials. For MVC assessment, participants were instructed to maximally contract their non-dominant leg for 5 s. Two trials, separated by 60 s, were completed prior to the submaximal fatiguing task. Visual feedback relating to force production and elapsed time was presented on a computer monitor throughout the MVC trials. Scripted verbal encouragement was also provided by the researcher. Peak torque (Nm) from the better of the two trials was taken as the MVC. Pre-tDCS MVC was used to calculate the target force for the fatiguing task.

Muscular endurance assessment

For the assessment of muscular endurance, a fatiguing task was performed whereby participants were required to maintain a submaximal isometric contraction at or above 30% of their pre-tDCS MVC force for as long as possible. Task failure was defined as an inability to produce force at or above the target for three consecutive seconds. To promote maximal performance while also limiting the potential for experimenter bias, participants received scripted verbal encouragement throughout the task. Participants also received visual feedback on a computer monitor indicating current and target force. No feedback regarding elapsed-time was given to participants during the fatiguing task.

Direct current stimulation

To ensure participant safety, the application of HD-tDCS adhered to the procedural recommendations of Villamar et al. (2013). A HD-tDCS multi-channel stimulation interface (Model 4X1-C2, Soterix Medical, New York, NY) attached to a 1 × 1 low-intensity direct current stimulator (Model 1300, Soterix Medical) delivered 20 min of HD-tDCS. Following the International 10/10 electroencephalography system, electrodes were placed in a 4 × 1 ring configuration with the centre electrode positioned on the scalp over the hand motor cortex contralateral to the non-dominant side (C3/C4) and return electrodes positioned in a ring around the centre anode at a radius of approximately 5 cm corresponding to Cz, F3/F4, T7/T8 and P3/P4. Electrodes were held in place using an electroencephalography recording cap. Both electrode placement and stimulation intensity replicated the methods used in previous research showing enhanced CPM following HD-tDCS (Flood, Waddington & Cathcart, 2016). Electrode placement over the hand motor cortex was also chosen to ensure that any functional change in the performance of the leg motor tasks was not due to increased cortico-spinal excitability of the involved limb. HD-Explore software (Version 2.3, Soterix Medical, New York, NY) was used to confirm the focality of the HD-tDCS induced electrical field (Fig. 1). For comparison, computational modelling for the conventional 35 m² sponge-electrodes is also presented (Fig. 1).

Participants received both active and sham (placebo) stimulation conditions. The active condition began with an initial 10 s ramping up period, where stimulation intensity was gradually increased from 0 mA to the target intensity of 2 mA. The current then remained at 2 mA for 20 min, followed by a gradual (10 s) ramping down to 0 mA.

Protocols for the administration of sham stimulation followed those of the active condition. However, in the sham condition, stimulation intensity ramped up to 2 mA and then immediately back down to 0 mA at the start and again at the end of the 20 min period. Previous research has supported the use of similar sham protocols as successful methods for the blinding of participants to their experimental condition (Borckardt et al., 2012). Protocols for the delivery of active and sham stimulation, including ramping periods, were automated by the stimulation device, ensuring consistency throughout the sample.

Motivation

The scale developed by Matthews, Campbell & Falconer (2001) was used for the assessment of motivation relating to the physical performance tasks. This scale measures success and intrinsic motivation for an upcoming task. All items are scored according to a five-point Likert scale. Higher total scores represent greater levels of motivation for both domains of motivation.

Procedure

Study protocols followed those approved by the local human research ethics committee (Project: 15-261). Prior to testing, written informed consent was obtained. Participants attended two testing sessions separated by at least 1 wk, where they received active or sham stimulation in a randomised counterbalanced order. A single-blind design was adopted, with participants unaware of the ordering of the stimulation conditions.

Motivation, pain inhibition, MVC, and muscular endurance were assessed before and after the HD-tDCS intervention. The calculation of the target force for both pre- and post-tDCS muscular endurance trials was based on pre-tDCS MVC performance. Pre- and post-tDCS physical performance measures were separated by 1 h. During this period, participants rested for 20 min, then underwent 20 min of HD-tDCS followed by post-tDCS motivation and CPM assessment. An additional 4 min rest period separated MVC and endurance trials. See Fig. 2 for schematic description of methods.

Data analysis

Data analysis was performed using SPSS statistical software (Version 21, IBM Corp, Armonk, NY). For all analyses, p < .050 was considered statistically significant. Means and standard deviations for all dependent variables are presented in Table 1. In accordance with prescribed publication guidelines (APA, 2010), significance values (p-values) and effect sizes (Cohen’s d or partial η²) are presented for all analyses.

Prior to analysis, assumptions for parametric tests were assessed and confirmed. Paired samples t-tests were used to assess differences in pre-tDCS motivation, pre-conditioning PPT, CPM, MVC and muscular endurance between the two sessions. Two-way repeated measures ANOVAs examined the effect of anodal and sham stimulation on MVC, muscular endurance, pre-conditioning PPT, CPM and motivation with the main factors of time (pre- and post-tDCS) and condition (active and sham). Normalised values were also calculated for peak MVC force (MVC-norm) and muscular endurance (endurance-norm) as (post-tDCS–pre-tDCS)/pre-tDCS. Therefore, these normalised values represent the relative (percentage) change in performance from pre- to post-tDCS, with negative values indicating a reduction in performance. Paired samples t-tests compared MVC-norm and endurance-norm across the two conditions.

Baseline assessment

Statistical analysis confirmed that pre-tDCS success motivation, t(11) = .80, p = .441, d = .17, intrinsic motivation, t(11) = .82, p = .429, d = .16, pre-conditioning PPT, t(11) =  − .63, p = .541, d = .16, CPM, t(11) = 1.07, p = .307, d = .52, MVC, t(11) = 1.68, p = .121, d = .20 and muscular endurance, t(11) =  − 1.75, p = .109, d = .32 did not differ between the two sessions.

Motivation

Self-reported success and intrinsic motivation (Table 1) exceeded the levels observed in previous research (Marcora, Staiano & Manning, 2009). Statistical analysis revealed that levels of success motivation did not differ over time, F(1, 11) = 1.77, p = .211, partial η² = .14, or condition, F(1, 11) = 2.39, p = .150, partial η² = .18. Similarly, levels of intrinsic motivation did not differ over time, F(1, 11) = 2.93, p = .115, d = .21, or condition F(1, 11) = 2.71, p = .128, d = .20. There was also no significant time × condition interaction effect for success motivation, F(1, 11) = 1.92, p = .194, partial η² = .15, or intrinsic motivation, F(1, 11) = .92, p = .358, d = .08, indicating that HD-tDCS did not affect participant motivation.

Pain measures

Pre-conditioning PPTs did not differ significantly as a function of time, F(1, 11) = .15, p = .708, d = .01, or condition, F(1, 11) = .02, p = .893, d = .00. The time × condition interaction effect for pre-conditioning PPT was also non-significant, F(1, 11) = 1.06, p = .324, d = .09.

Although there was no significant main effect of time, F(1, 11) = .21, p = .654, partial η² = .02, or condition, F(1, 11) = .31, p = .588, partial η² = .03, on CPM, a statistically significant interaction effect was uncovered, F(1, 11) = 5.07, p = .046, partial η² = .32, demonstrating increased endogenous pain inhibition in the active condition relative to the sham condition.

Physical performance

For MVC force, statistical analysis revealed a significant main effect for time, F(1, 11) = 9.61, p = .010, partial η² = .47, but not condition, F(1, 11) = .72, p = .415, partial η² = .06. There was also no significant time × condition interaction for MVC performance, F(1, 11) = 2.89, p = .117, partial η² = .21. In addition, MVC-norm did not differ significantly between active (−11.24 ± 9.42%) and sham (−.31 ± 15.22%) conditions, t(11) = 1.91, p = .083, d =  − .89.

Muscular endurance decreased from pre- to post-tDCS irrespective of condition, F(1, 11) = 5.75, p = .035, partial η² = .34. However, there was no significant main effect for condition, F(1, 11) = 2.53, p = .140, partial η² = .19. The reduction in endurance performance from pre- to post-sham stimulation (M = 23.15 ± 37.67 s reduction) was greater than the reduction from pre- to post-active stimulation (M = 11.58 ± 15.81 s reduction). When expressed as a relative change from pre- to post-tDCS, endurance time decreased by 13.30 ± 24.56% following sham stimulation, but only 9.39 ± 12.55% following active. However, both the relative, t(11) = .68, p = .511, d = .21, and absolute, F(1, 11) = 1.95, p = .190, partial η² = .15, performance advantages gained from active HD-tDCS were not deemed statistically significant.

Limitations

Several limitations of the current study should be noted. The endurance performance task required participants to remain focussed on the computer-displayed force trace and ensure that their force production stayed above the target. Within-task pain assessment was not included as the additional cognitive burden on the participant may have produced a sub-maximal performance. Therefore, as discussed above, we cannot say with certainty that the observed relative increase in endogenous pain inhibitory capacity resulting from active HD-tDCS did reduce exercise-induced pain perception. Future research should attempt to address this limitation through the inclusion of pain assessment both at rest and during the performance of the exercise tasks.

It has been widely reported that the performance of physical activity causes acute widespread hypoalgesia (Koltyn, 2000; Koltyn, 2002; Drury et al., 2004). It is possible that pre-tDCS physical performance may have impacted upon post-tDCS pain assessment. However, the exercise-induced hypoalgesic effect decays quickly after the completion of the exercise task (Koltyn, 2000). Therefore, to account for the potential impact of exercise-induced hypoalgesia, pre-tDCS muscular endurance trials and post-tDCS pain assessment were separated by 40 min (Fig. 2). This rest-period was consistent across the two experimental conditions. Statistical analysis confirmed that pain sensitivity returned to resting levels, with pre-conditioning PPTs not differing between pre- and post-tDCS measures.

The spatial specificity of the electrical current applied via HD-tDCS is greater than that of conventional tDCS (Datta et al., 2009). Computational modelling confirmed this assumption for the current montage (Fig. 1). These differences in the distribution of the stimulated neural region resulting from high-definition and conventional tDCS methods limit the potential for comparisons to be made between the current findings and the findings of previous research. In addition, although HD-tDCS allows for greater stimulation specificity, Fig. 1 suggests that cortical stimulation may have dispersed beyond the targeted hand motor cortex, potentially inducing unwanted neuronal changes in neighbouring regions. In particular, cortical stimulation may have induced changes in cortico-spinal excitability to the leg motor region. However, previous research has achieved somatotopically specific effects using conventional tDCS (Tanaka et al., 2009), suggesting that the more targeted form of cortical stimulation utilised in the current study was sufficiently localised to the target region. To further clarify the effects of targeted cortical stimulation on motor function, we encourage the examination of tDCS-induced changes in cortico-spinal excitability using methods such as transcranial magnetic stimulation.

Caution should be taken when generalising the current findings to the wider population. The current study sampled only young, pain-free males. Research suggests that pain inhibitory capacity is reduced with age (Grashorn et al., 2013), and that males possess increased pain inhibitory responses (Popescu et al., 2010). In addition, chronic pain sufferers exhibit maladaptive pain inhibition (Lautenbacher & Rollman, 1997). It is possible then that in these populations, the effects of HD-tDCS on pain inhibition and exercise performance may be different to the effects seen in the current study.

Similarly, task duration is an important factor to consider in the generalisation of the current results. The mean endurance time across all conditions was approximately 100 s. The factors influencing performance regulation are argued to be task-specific, particularly dependent on the duration of the task (Abbiss & Laursen, 2008). It is possible then that the role of endogenous pain inhibition may be more pronounced in longer duration tasks. Indeed, increased cycling performance following acetaminophen ingestion occurred in endurance tasks exceeding 20 min (Mauger, Jones & Williams, 2010). Future research should continue to examine the potential mediating role of task duration on the factors regulating exercise performance.

No habituation trials were conducted prior to the experimental sessions to ensure a consistency in endurance performance prior to the intervention. The novelty of the performance task may have resulted in the natural performance variations diluting the performance effects of the HD-tDCS intervention. However, statistical analysis demonstrated no significant difference in pre-tDCS endurance time or pre-tDCS MVC force over the two sessions. This suggests that the examination of the performance effects of HD-tDCS were conducted from a stable baseline muscular endurance and MVC assessment.

The current study adopted a single blinded experimental design. In order to minimise the potential for experimenter bias, standardised protocols were implemented for the measurement of both pain and exercise performance. For PPT assessment, rate of pressure application (1 kg/s) and the inter-stimulus interval (30 s) remained consistent for all participants. The rate of cuff inflation for occlusion was also consistent (∼20 mmHg/s). To limit the potential for experimenter bias in the exercise performance tasks, performance feedback displayed on a computer monitor was consistent across sessions. Scripted verbal encouragement during the MVC and endurance tasks also accounted for potential experimenter bias.

There is currently a lack of consistency in the CPM methods used for the assessment of pain modulation capacity (Yarnitsky et al., 2015). PPT assessment as the TS in the CPM protocol is widely utilised, however, several alternate methods such as noxious heat are also used (Pud, Granovsky & Yarnitsky, 2009). Methodological inconsistencies are also evident in the methods used for the CS. Following previous research methodologies (Cathcart et al., 2009), cuff occlusion was used to inhibit the perceived intensity of the TS. To ensure that the perceived intensity of the CS was consistent for all participants, the actual intensityand duration of cuff occlusion varied. It is possible that these variations in the actual intensity and duration of the CS produced varied inhibitory responses. Due to these methodological inconsistencies, we urge caution when generalising the magnitude of pain inhibition observed in the current study.